Computer-readable recording medium storing information processing program, information processing method, and information processing device

Abstract

A non-transitory computer-readable recording medium storing an information processing program for causing a computer to execute processing includes: extracting partial data that corresponds to each of a plurality of patterns of which an index according to an appearance frequency in a plurality of training samples is equal to or more than a second threshold, that is a pattern of a combination of one or more feature amounts of which a contribution degree to estimation of a machine learning model is equal to or more than a first threshold; calculating a likelihood of an estimation result of a partial model for each pattern, in a case where the partial data extracted from the estimation target data is input to a corresponding partial model among the partial models trained for the respective patterns; and outputting the partial data selected based on the likelihood calculated for each partial model for each pattern.

What is claimed is: 
     
         1 . A non-transitory computer-readable recording medium storing an information processing program for causing a computer to execute processing comprising:
 extracting partial data that corresponds to each of a plurality of patterns of which an index according to an appearance frequency in a plurality of training samples is equal to or more than a second threshold, that is a pattern of a combination of one or more feature amounts of which a contribution degree to estimation of a machine learning model is equal to or more than a first threshold, from among the plurality of training samples that includes the plurality of feature amounts, from estimation target data that includes a plurality of feature amounts, to be a target of estimation processing by using the machine learning model;   calculating a likelihood of an estimation result of a partial model for each pattern, in a case where the partial data extracted from the estimation target data is input to a corresponding partial model among the partial models trained for the respective patterns; and   outputting the partial data selected based on the likelihood calculated for each partial model for each pattern, as an estimation basis of the machine learning model for the estimation target data.   
     
     
         2 . The non-transitory computer-readable recording medium according to  claim 1 , for causing the computer to execute processing further comprising: specifying the plurality of patterns from each of the plurality of training samples and training a partial model for each pattern, by using a partial training sample obtained by extracting partial data that corresponds to each of the plurality of specified patterns. 
     
     
         3 . The non-transitory computer-readable recording medium according to  claim 2 , wherein the training the partial model includes extracting partial data that corresponds to a new pattern from a plurality of new training samples and training a partial model that corresponds to the new pattern, until accuracy of an estimation result of a new machine learning model trained by using the plurality of new training samples obtained by removing the partial data used to train the partial model from each of the plurality of training samples falls below a third threshold. 
     
     
         4 . The non-transitory computer-readable recording medium according to  claim 1 , wherein the selecting the partial data based on the likelihood includes selecting the partial data input to the partial model of which the likelihood is equal to or more than a fourth threshold. 
     
     
         5 . The non-transitory computer-readable recording medium according to  claim 1 , wherein the outputting the estimation basis includes aggregating and outputting a plurality of pieces of the partial data selected based on the likelihood. 
     
     
         6 . The non-transitory computer-readable recording medium according to  claim 1 , wherein the training sample and the estimation target data are graph data that represents the plurality of feature amounts with a plurality of nodes and an edge that couples between the nodes, the partial data is a partial graph of the graph data that indicates the training sample or the estimation target data, and the machine learning model and the partial model are machine learning models capable of handling the graph data. 
     
     
         7 . An information processing method, in which a computer executes processing comprising:
 extracting partial data that corresponds to each of a plurality of patterns of which an index according to an appearance frequency in a plurality of training samples is equal to or more than a second threshold, that is a pattern of a combination of one or more feature amounts of which a contribution degree to estimation of a machine learning model is equal to or more than a first threshold, from among the plurality of training samples that includes the plurality of feature amounts, from estimation target data that includes a plurality of feature amounts, to be a target of estimation processing by using the machine learning model;   calculating a likelihood of an estimation result of a partial model for each pattern, in a case where the partial data extracted from the estimation target data is input to a corresponding partial model among the partial models trained for the respective patterns; and   outputting the partial data selected based on the likelihood calculated for each partial model for each pattern, as an estimation basis of the machine learning model for the estimation target data.   
     
     
         8 . An information processing device comprising:
 a memory; and   a processor coupled to the memory, the processor being configured to perform processing comprising:   extracting partial data that corresponds to each of a plurality of patterns of which an index according to an appearance frequency in a plurality of training samples is equal to or more than a second threshold, that is a pattern of a combination of one or more feature amounts of which a contribution degree to estimation of a machine learning model is equal to or more than a first threshold, from among the plurality of training samples that includes the plurality of feature amounts, from estimation target data that includes a plurality of feature amounts, to be a target of estimation processing by using the machine learning model;   calculating a likelihood of an estimation result of a partial model for each pattern, in a case where the partial data extracted from the estimation target data is input to a corresponding partial model among the partial models trained for the respective patterns; and   outputting the partial data selected based on the likelihood calculated for each partial model for each pattern, as an estimation basis of the machine learning model for the estimation target data.